Obstacle Avoidance Distributed Tracking of Networked UAVs with Online Path Planning

This paper addresses the dual challenges of obstacle avoidance and distributed tracking in networked Unmanned Aerial Vehicles (UAVs) by leveraging online planning techniques. Specifically, we introduce a trajectory planning framework for the leader UAV, which is based on the Ego-Planner and utilizes high-performance sensors alongside an advanced onboard computing system. The optimized trajectory of the leader UAV serves as a dynamic reference for the entire UAV network. Each low-cost follower UAV implements an online learning algorithm to extract high-order trajectory information from the leader, enabling the generation of feedforward control components. These feedforward components are integrated into a feedbackfeedforward control architecture, facilitating precise distributed tracking and effective obstacle avoidance. The proposed method significantly enhances cooperative UAV navigation in complex and obstacle-dense environments, demonstrating that low-cost UAVs can achieve reliable performance. Extensive simulations validate the approach's effectiveness, showing superior obstacle avoidance and tracking accuracy compared to conventional methods.